﻿664 Harris: Distribution and correlation of sexe 



metical error. As will be shown presently the correlatioi 

 negative but positive in sign, though numerically small, 

 details concerning habitats and other biological conditi< 

 original papers musi 



II. Statistical Analysis of Data 

 Consider first the means and variabilities of the number of 

 staminate and pistillate flowers in the inflorescences of the two 

 species. For this discussion only the total materials collected 

 for each species will be used. In A. vulgare the range of variation 

 in number of staminate flowers is very great. The classes and the 

 frequency with which they occur, together with the total number of 

 pistillate flowers found on the inflorescences, are shown in Table I. 



UENCY OF Staminate ] 



Pistillate Flowe 



813 35 41 ! 209 



16 I 13 37 37 i 254 



17 I I 38 30 ! 189 



18 3 12 39 39 ■ 263 



20 I 4 41 26 I 175 



21 4 13 42 21 \ 147 



25 9 , 50 46 : 5 I 38 



27 14 64 48 4 1 ^9 



The vananon m tnc number of pistillate flowers may be calculated 

 from the totals of Table 1 1 . The data for A . proboscideum appear 

 in the totals of Table III. From these I deduce the constants for 

 number of flowers per inflorescence given in Table IV. 



The fact that the two species differ greatly in the numbers of 

 the two types of flowers which they produce is well known to taxon- 



